Behavioral Targeting: Wanted by Everyone, Understood By Few
Behavioral targeting firms promise to deliver consumers who are more likely to be interested in a particular product or message, based on their Web viewing habits, to advertisers. But the firms themselves haven't yet determined a standard for the behavior pattern that transforms Web viewers into appropriate targets.
Nevertheless, said the executives, some advertisers accept and buy the behavioral segments sold to them by publishers at face value, without questioning the data that's used to generate given audience segments.
Cameron Yuill, vice president of partner marketing for Cendant, noted that Cendant's smaller advertisers often are excited about reaching a highly targeted number of people, but have no idea what the segments are comprised of--which in turn handicaps negotiations.
Generally, behavioral targeting firms look at people's viewing habits on a particular publisher's Web site and then send those people ads based on the types of pages they've visited. For instance, a viewer who visits the automobile section of a local newspaper might be deemed a potential car buyer; he might then receive ads directed at automobile purchasers wherever he goes on the site.
Dave Morgan of Tacoda Systems champions the idea of segmenting an audience into marketing categories, such as potential car buyers, based on exactly such criteria as reading articles about cars online.
But Omar Tawakol, senior vice president of marketing of rival behavioral targeting company Revenue Science, says that simply placing users in categories is inadequate unless the quality of a publisher's audience is also considered.
For example, said Tawakol, a Web user identified as being in the market for a car, based on viewing behavior within the site Edmunds.com--which offers advice for those looking to purchase cars--probably has exhibited far more in-market car-buying characteristics than the local paper's reader, who may be simply a passive enthusiast. The frequency of behavior and the quality of behavior exhibited at each site are different, yet both users might qualify as in-market car buyers under a minimum naming standard.
But not all publishers agree with Revenue Science that audience quality should be taken into consideration when establishing standards. Reuters.com vice president of media sales Walker Jacobs said that evaluating the quality of a given segment "is a media planning job," as opposed to a task for behavioral marketers.
Walker added that advertisers and publishers need to be clear about how they're classifying people into segments. "For the most part, my investor segment is going to be different from an advertiser's idea of an investor segment--it's not necessarily an apples-to-apples comparison," Jacobs said.
The panelists agreed that advertisers almost always want more granular segment definitions, while publishers push for broader criteria. For example, publishers might try to sell a segment of auto buyers that looked at an auto page once in the last two months as "in-market" precisely because it generates a greater reach. "Publishers are very good at putting marketing spin on this stuff," Jacobs said.
Publishers sometimes add to the confusion by not including segment definitions in their media kits. Shauna Monkman, FT.com's global head of online sales, noted that FT.com provides advertisers with a list of available segments, but does not include definitions unless they're specifically requested. Many advertisers, she said, either don't know or don't care enough to ask.
For the savvy advertisers that request more information, most publishers are willing to create custom segments, although Reuters.com's Jacobs scoffed at the notion. "As soon as we try to be all things to all people we dilute our brand," he said.